Scaling a UX Org from 30 to 450, with Shopify’s former Head of Design Kyle Peatt

Kyle Peatt says his first design work was creating forum tags and other graphics for his Counter-Strike clan. A friend interested in web development pulled him toward digital design, and in high school they built websites for local businesses. After the dot-com crash made technology seem like an implausible career, Kyle studied arts for a year, then attended Calgary’s Southern Alberta Institute of Technology because he could not draw well enough for a traditional design-school portfolio. The technical program taught him print production, giving him a production-design foundation alongside his self-taught digital skills.

While studying and taking summer design jobs, Kyle joined an SMS-marketing startup as a product designer, building mobile websites before the iPhone. He worked on sites for restaurants and a Kung Fu Panda promotion where fruit carried codes that users texted to receive a mobile URL. His most memorable project was a UFC site built with one developer for WAP devices such as Motorola Razrs. They reorganized fight statistics for tiny screens and launched on fight night, but traffic logs showed users mainly viewed pictures of the “fight girls.” Kyle calls it an early lesson in the gap between designing and observing real user behavior.

Kyle moved to Vancouver startup Mobify and spent just over three years there, his first opportunity to think deliberately about team design rather than only his own work. He joined when the company had about 10 people and one other designer, then grew the professional-services group from two designers plus front-end engineers into a focused team of roughly 20: about 10 designers and nine front-end developers. He intentionally combined design and front-end development because he found it difficult to separate how something is built from what is being built. Mobify became the place where he first formed strong opinions about scaling design teams.

Feeling that he had plateaued at Mobify, Kyle interviewed with Shopify despite living across the country in Vancouver and knowing little about the product. He was attracted by the company’s design-founder influence: Daniel’s human-centered values appeared not only in Shopify’s product but in how employees worked. Shopify also offered design leaders and a co-founder from whom Kyle could test opinions formed without a previous design mentor. He joined the week Shopify went public, when roughly 30 designers across three offices reported to only two people. Six years later, the UX organization had grown to about 450 people.

When Kyle joined, Shopify had only three or four front-end developers, and they were not necessarily called engineers. The group, previously led by BEM creator John Snook, built reusable Rails components that let engineers across the company build the product rapidly; much of the interface was server-rendered. Kyle and Alan note that this structure differed sharply from the later UX organization and from conventional assumptions about design and engineering. The rapid growth after Shopify went public forced leaders to decide whether to preserve the hyperflat, pooled model or create clearer problem-oriented teams.

Kyle initially led five designers assigned across different Shopify problem areas. One area, Home, was widely described as impossible because stakeholders disagreed and the designer had to navigate conflicting opinions. Kyle immersed himself in that crisis and helped the designer cut through the obstacles, but then realized the other four designers had received little support. The experience showed him that assigning one design lead to many people and problems makes the most urgent fire consume the lead’s attention. He began shifting from people-oriented leadership toward design leads who were deeply accountable to specific problem spaces.

Kyle proposed pairing two designers with every problem area rather than assigning one designer to each area while one lead covered five problems. Each lead would then focus on roughly two or three problems, with the eventual goal of one design lead aligned to each product manager, matching Shopify’s existing engineering structure. The pair could have complementary strengths—such as visual and product or experience design—and one experienced designer could help onboard another. Kyle also argued that resource allocation should begin with the largest business and user outcomes, not an even “peanut butter” spread across every dashboard or feature.

Kyle defines product management as owning business outcomes and strategic alignment, not dictating an interface. For example, product could identify the opportunity to let Shopify merchants buy and attach shipping labels inside the product, while design would discover the best user outcome and solution with engineering. The three disciplines must share deep context: engineering must understand design intent, design must understand product intent, and each person should be able to represent the others’ interests when they are absent. Kyle warns that the triad fails when one leg dominates—for example, when product has final say or engineering’s implementation constraints determine everything.

Kyle eventually implemented a structure in which designers, researchers, content strategists, and specialized front-end developers were pulled from small disciplinary pools into UX teams organized around problems. Each team had a UX lead accountable for the experience, with dedicated expertise rather than researchers or content strategists parachuting into projects without context. This meant a designer might report to a UX lead whose background was content strategy, a contentious change because that lead was responsible for both management and the quality of the problem-solving. Staff designers and equivalent technical leaders remained embedded in problem areas, mentoring peers and improving practice rather than becoming detached specialists.

Kyle traces his view of design opinion to an old party argument with Alan about macOS’s green title-bar button: Kyle wanted a setting for full-screen versus constrained windows, while Alan argued that Apple should choose a good default rather than expose every preference. Kyle now defines opinion as experience, research, data, and taste combined, not an arbitrary instinct. Designers should state strong opinions but hold them weakly enough to change when evidence changes. He says the goal is neither to surrender to research or metrics nor to impose personal taste; it is to understand why an opinion exists and use that reasoning to improve it.